This action might not be possible to undo. Are you sure you want to continue?
©Silberschatz, Korth and Sudarshan 7.2 Database System Concepts
Chapter 7: Relational Database Design Chapter 7: Relational Database Design
First Normal Form
Pitfalls in Relational Database Design
Functional Dependencies
Decomposition
BoyceCodd Normal Form
Third Normal Form
Multivalued Dependencies and Fourth Normal Form
Overall Database Design Process
©Silberschatz, Korth and Sudarshan 7.3 Database System Concepts
First Normal Form First Normal Form
Domain is atomic if its elements are considered to be indivisible units
Examples of nonatomic domains:
Set of names, composite attributes
Identification numbers like CS101 that can be broken up into parts
A relational schema R is in first normal formif the domains of all attributes of R
are atomic
Nonatomic values complicate storage and encourage redundant (repeated)
storage of data
E.g. Set of accounts stored with each customer, and set of owners stored with
each account
We assume all relations are in first normal form (revisit this in Chapter 9 on Object
Relational Databases)
©Silberschatz, Korth and Sudarshan 7.4 Database System Concepts
First Normal Form (Contd.) First Normal Form (Contd.)
Atomicity is actually a property of how the elements of the domain are used.
E.g. Strings would normally be considered indivisible
Suppose that students are given roll numbers which are strings of the form
CS0012 or EE1127
If the first two characters are extracted to find the department, the domain of roll
numbers is not atomic.
Doing so is a bad idea: leads to encoding of information in application program
rather than in the database.
©Silberschatz, Korth and Sudarshan 7.5 Database System Concepts
Pitfalls in Relational Database Design Pitfalls in Relational Database Design
Relational database design requires that we find a ³good´ collection
of relation schemas. A bad design may lead to
Repetition of Information.
Inability to represent certain information.
Design Goals:
Avoid redundant data
Ensure that relationships among attributes are represented
Facilitate the checking of updates for violation of database integrity
constraints.
©Silberschatz, Korth and Sudarshan 7.6 Database System Concepts
Example Example
Consider the relation schema:
Lendingschema = (branchname, branchcity, assets,
customername, loannumber, amount)
Redundancy:
Data for branchname, branchcity, assets are repeated for each loan that a branch makes
Wastes space
Complicates updating, introducing possibility of inconsistency of assets value
Null values
Cannot store information about a branch if no loans exist
Can use null values, but they are difficult to handle.
©Silberschatz, Korth and Sudarshan 7.7 Database System Concepts
Decomposition Decomposition
Decompose the relation schema Lendingschema into:
Branchschema = (branchname, branchcity,assets)
Loaninfoschema = (customername, loannumber,
branchname, amount)
All attributes of an original schema (R) must appear in the
decomposition (R
1
, R
2
):
R = R
1
R
2
Losslessjoin decomposition.
For all possible relations r on schema R
r = 
R1
(r) 
R2
(r)
©Silberschatz, Korth and Sudarshan 7.8 Database System Concepts
Example of Non Lossless Example of Non LosslessJoin Decomposition Join Decomposition
Decomposition of R = (A, B)
R
2
= (A) R
2
= (B)
A B
¬
¬
¡
1
2
1
A
¬
¡
B
1
2
r

A
(r)

B(r)

A
(r) 
B
(r)
A B
¬
¬
¡
¡
1
2
1
2
©Silberschatz, Korth and Sudarshan 7.9 Database System Concepts
Goal Goal ³³Devise a Theory for the Following Devise a Theory for the Following
Decide whether a particular relation R is in ³good´ form.
In the case that a relation R is not in ³good´ form, decompose it into a set of
relations {R
1
, R
2
, ..., R
n
} such that
each relation is in good form
the decomposition is a losslessjoin decomposition
Our theory is based on:
functional dependencies
multivalued dependencies
©Silberschatz, Korth and Sudarshan 7.10 Database System Concepts
Functional Dependencies Functional Dependencies
Constraints on the set of legal relations.
Require that the value for a certain set of attributes determines uniquely the
value for another set of attributes.
A functional dependency is a generalization of the notion of a key.
©Silberschatz, Korth and Sudarshan 7.11 Database System Concepts
Functional Dependencies (Cont.) Functional Dependencies (Cont.)
Let R be a relation schema
¬ _ R and ¡ _ R
The functional dependency
¬ ÷¡
holds on R if and only if for any legal relations r(R), whenever any two tuples t
1
and t
2
of r agree on the attributes ¬, they also agree on the attributes ¡. That is,
t
1
[¬] = t
2
[¬] ÷ t
1
[¡ ] = t
2
[¡ ]
Example: Consider r(A,B) with the following instance of r.
On this instance, A ÷B does NOT hold, but B ÷A does hold.
1 4
1 5
3 7
©Silberschatz, Korth and Sudarshan 7.12 Database System Concepts
Functional Dependencies (Cont.) Functional Dependencies (Cont.)
K is a superkey for relation schema R if and only if K ÷R
K is a candidate key for R if and only if
K ÷R, and
for no ¬ · K, ¬ ÷R
Functional dependencies allow us to express constraints that cannot be
expressed using superkeys. Consider the schema:
Loaninfoschema = (customername, loannumber,
branchname, amount).
We expect this set of functional dependencies to hold:
loannumber ÷amount
loannumber ÷branchname
but would not expect the following to hold:
loannumber ÷customername
©Silberschatz, Korth and Sudarshan 7.13 Database System Concepts
Use of Functional Dependencies Use of Functional Dependencies
We use functional dependencies to:
test relations to see if they are legal under a given set of functional dependencies.
If a relation r is legal under a set F of functional dependencies, we say that r
satisfies F.
specify constraints on the set of legal relations
We say that F holds on R if all legal relations on R satisfy the set of functional
dependencies F.
Note: A specific instance of a relation schema may satisfy a functional
dependency even if the functional dependency does not hold on all legal
instances. For example, a specific instance of Loanschema may, by chance,
satisfy
loannumber ÷customername.
©Silberschatz, Korth and Sudarshan 7.14 Database System Concepts
Functional Dependencies (Cont.) Functional Dependencies (Cont.)
A functional dependency is trivial if it is satisfied by all instances of a relation
E.g.
customername, loannumber ÷customername
customername ÷customername
In general, ¬ ÷¡ is trivial if ¡ _ ¬
©Silberschatz, Korth and Sudarshan 7.15 Database System Concepts
Closure of a Set of Functional Dependencies Closure of a Set of Functional Dependencies
Given a set F set of functional dependencies, there are certain other functional
dependencies that are logically implied by F.
E.g. If A ÷B and B ÷C, then we can infer that A ÷C
The set of all functional dependencies logically implied by F is the closure of F.
We denote the closure of F by F
+
.
We can find all of F
+
by applying Armstrong¶s Axioms:
if ¡ _ ¬, then ¬ ÷¡ (reflexivity)
if ¬ ÷¡, then K ¬ ÷ K ¡ (augmentation)
if ¬ ÷¡, and ¡ ÷K, then ¬ ÷ K (transitivity)
These rules are
sound (generate only functional dependencies that actually hold) and
complete (generate all functional dependencies that hold).
©Silberschatz, Korth and Sudarshan 7.16 Database System Concepts
Example Example
R = (A, B, C, G, H, I)
F = { A ÷B
A ÷C
CG ÷H
CG ÷I
B ÷H}
some members of F
+
A ÷H
by transitivity from A ÷B and B ÷H
AG ÷I
by augmenting A ÷C with G, to get AG ÷CG
and then transitivity with CG ÷I
CG ÷HI
from CG ÷H and CG ÷I : ³union rule´ can be inferred from
± definition of functional dependencies, or
± Augmentation of CG ÷I to infer CG ÷CGI, augmentation of
CG ÷H to infer CGI ÷HI, and then transitivity
©Silberschatz, Korth and Sudarshan 7.17 Database System Concepts
Procedure for Computing F Procedure for Computing F
+ +
To compute the closure of a set of functional dependencies F:
F
+
= F
repeat
for each functional dependency f in F
+
apply reflexivity and augmentation rules on f
add the resulting functional dependencies to F
+
for each pair of functional dependencies f
1
and f
2
in F
+
if f
1
and f
2
can be combined using transitivity
then add the resulting functional dependency to F
+
until F
+
does not change any further
NOTE: We will see an alternative procedure for this task later
©Silberschatz, Korth and Sudarshan 7.18 Database System Concepts
Closure of Functional Dependencies (Cont.) Closure of Functional Dependencies (Cont.)
We can further simplify manual computation of F
+
by using the following
additional rules.
If ¬ ÷¡ holds and ¬ ÷K holds, then ¬ ÷¡ K holds (union)
If ¬ ÷¡ K holds, then ¬ ÷¡ holds and ¬ ÷K holds (decomposition)
If ¬ ÷¡ holds and K ¡ ÷o holds, then ¬ K ÷o holds
(pseudotransitivity)
The above rules can be inferred from Armstrong¶s axioms.
©Silberschatz, Korth and Sudarshan 7.19 Database System Concepts
Closure of Attribute Sets Closure of Attribute Sets
Given a set of attributes ¬, define the closure of ¬ under F (denoted by ¬
+
) as
the set of attributes that are functionally determined by ¬ under F:
¬ ÷¡ is in F
+
¡ _ ¬
+
Algorithm to compute ¬
+
, the closure of ¬ under F
result := ¬;
while (changes to result) do
for each ¡ ÷K in F do
begin
if ¡ _ result then result := result K
end
©Silberschatz, Korth and Sudarshan 7.20 Database System Concepts
Example of Attribute Set Closure Example of Attribute Set Closure
R = (A, B, C, G, H, I)
F = {A ÷B
A ÷C
CG ÷H
CG ÷I
B ÷H}
(AG)
+
1. result = AG
2. result = ABCG (A ÷C and A ÷B)
3. result = ABCGH (CG ÷H and CG _ AGBC)
4. result = ABCGHI (CG ÷I and CG _ AGBCH)
Is AG a candidate key?
1. Is AG a super key?
1. Does AG ÷R?
2. Is any subset of AG a superkey?
1. Does A
+
÷R?
2. Does G
+
÷R?
©Silberschatz, Korth and Sudarshan 7.21 Database System Concepts
Uses of Attribute Closure Uses of Attribute Closure
There are several uses of the attribute closure algorithm:
Testing for superkey:
To test if ¬ is a superkey, we compute ¬
+,
and check if ¬
+
contains all attributes of
R.
Testing functional dependencies
To check if a functional dependency ¬ ÷¡ holds (or, in other words, is in F
+
), just
check if ¡ _ ¬
+
.
That is, we compute ¬
+
by using attribute closure, and then check if it contains ¡.
Is a simple and cheap test, and very useful
Computing closure of F
For each K _ R, we find the closure K
+
, and for each S _ K
+
, we output a functional
dependency K ÷S.
©Silberschatz, Korth and Sudarshan 7.22 Database System Concepts
Canonical Cover Canonical Cover
Sets of functional dependencies may have redundant dependencies that can
be inferred from the others
Eg: A ÷C is redundant in: {A ÷B, B ÷C, A ÷C}
Parts of a functional dependency may be redundant
E.g. on RHS: {A ÷B, B ÷C, A ÷CD} can be simplified to
{A ÷B, B ÷C, A ÷D}
E.g. on LHS: {A ÷B, B ÷C, AC ÷D} can be simplified to
{A ÷B, B ÷C, A ÷D}
Intuitively, a canonical cover of F is a ³minimal´ set of functional dependencies
equivalent to F, with no redundant dependencies or having redundant parts of
dependencies
©Silberschatz, Korth and Sudarshan 7.23 Database System Concepts
Extraneous Attributes Extraneous Attributes
Consider a set F of functional dependencies and the functional dependency
¬ ÷¡ in F.
Attribute A is extraneous in ¬ if A ¬
and F logically implies (F ± {¬ ÷¡}) {(¬ ± A) ÷¡}.
Attribute A is extraneous in ¡ if A ¡
and the set of functional dependencies
(F ± {¬ ÷¡}) {¬ ÷(¡ ± A)} logically implies F.
Note: implication in the opposite direction is trivial in each of the cases
above, since a ³stronger´ functional dependency always implies a weaker
one
Example: Given F = {A ÷C, AB ÷C }
B is extraneous in AB ÷C because A ÷C logically implies
AB ÷C.
Example: Given F = {A ÷C, AB ÷CD}
C is extraneous in AB ÷CD since A ÷C can be inferred even after deleting
C
©Silberschatz, Korth and Sudarshan 7.24 Database System Concepts
Testing if an Attribute is Extraneous Testing if an Attribute is Extraneous
Consider a set F of functional dependencies and the functional dependency ¬
÷¡ in F.
To test if attribute A ¬ is extraneous in ¬
1. compute (A ± {¬})
+
using the dependencies in F
2. check that (A ± {¬})
+
contains ¬; if it does, A is extraneous
To test if attribute A ¡ is extraneous in ¡
1. compute ¬
+
using only the dependencies in
F¶ = (F ± {¬ ÷¡}) {¬ ÷(¡ ± A)},
2. check that ¬
+
contains A; if it does, A is extraneous
©Silberschatz, Korth and Sudarshan 7.25 Database System Concepts
Canonical Cover Canonical Cover
A canonical cover for F is a set of dependencies F
c
such that
F logically implies all dependencies in F
c,
and
F
c
logically implies all dependencies in F, and
No functional dependency in F
c
contains an extraneous attribute, and
Each left side of functional dependency in F
c
is unique.
To compute a canonical cover for F:
repeat
Use the union rule to replace any dependencies in F
¬
1
÷¡
1
and ¬
1
÷¡
1
with ¬
1
÷¡
1
¡
2
Find a functional dependency ¬ ÷¡ with an
extraneous attribute either in ¬ or in ¡
If an extraneous attribute is found, delete it from ¬ ÷¡
until F does not change
Note: Union rule may become applicable after some extraneous attributes have
been deleted, so it has to be reapplied
©Silberschatz, Korth and Sudarshan 7.26 Database System Concepts
Example of Computing a Canonical Cover Example of Computing a Canonical Cover
R = (A, B, C)
F = {A ÷BC
B ÷C
A ÷B
AB ÷C}
Combine A ÷BC and A ÷B into A ÷BC
Set is now {A ÷BC, B ÷C, AB ÷C}
A is extraneous in AB ÷C because B ÷C logically implies AB ÷C.
Set is now {A ÷BC, B ÷C}
C is extraneous in A ÷BC since A ÷BC is logically implied by A ÷B
and B ÷C.
The canonical cover is:
A ÷B
B ÷C
©Silberschatz, Korth and Sudarshan 7.27 Database System Concepts
Goals of Normalization Goals of Normalization
Decide whether a particular relation R is in ³good´ form.
In the case that a relation R is not in ³good´ form, decompose it into a set of
relations {R
1
, R
2
, ..., R
n
} such that
each relation is in good form
the decomposition is a losslessjoin decomposition
Our theory is based on:
functional dependencies
multivalued dependencies
©Silberschatz, Korth and Sudarshan 7.28 Database System Concepts
Decomposition Decomposition
Decompose the relation schema Lendingschema into:
Branchschema = (branchname, branchcity,assets)
Loaninfoschema = (customername, loannumber,
branchname, amount)
All attributes of an original schema (R) must appear in the decomposition (R
1
,
R
2
):
R = R
1
R
2
Losslessjoin decomposition.
For all possible relations r on schema R
r = 
R1
(r) 
R2
(r)
A decomposition of R into R
1
and R
2
is lossless join if and only if at least one of
the following dependencies is in F
+
:
R
1
R
2
÷R
1
R
1
R
2
÷R
2
©Silberschatz, Korth and Sudarshan 7.29 Database System Concepts
Example of Lossy Example of LossyJoin Decomposition Join Decomposition
Lossyjoin decompositions result in information loss.
Example: Decomposition of R = (A, B)
R
2
= (A) R
2
= (B)
A B
¬
¬
¡
1
2
1
A
¬
¡
B
1
2
r

A
(r)

B(r)

A
(r) 
B
(r)
A B
¬
¬
¡
¡
1
2
1
2
©Silberschatz, Korth and Sudarshan 7.30 Database System Concepts
Normalization Using Functional Dependencies Normalization Using Functional Dependencies
When we decompose a relation schema R with a set of functional
dependencies F into R
1
, R
2
,.., R
n
we want
Losslessjoin decomposition: Otherwise decomposition would result in information
loss.
No redundancy: The relations R
i
preferably should be in either BoyceCodd Normal
Form or Third Normal Form.
Dependency preservation: Let F
i
be the set of dependencies F
+
that include only
attributes in R
i
.
Preferably the decomposition should be dependency preserving, that is, (F
1
F
2
« F
n
)
+
= F
+
Otherwise, checking updates for violation of functional dependencies may require
computing joins, which is expensive.
©Silberschatz, Korth and Sudarshan 7.31 Database System Concepts
Example Example
R = (A, B, C)
F = {A ÷B, B ÷C)
R
1
= (A, B), R
2
= (B, C)
Losslessjoin decomposition:
R
1
R
2
= {B} and B ÷BC
Dependency preserving
R
1
= (A, B), R
2
= (A, C)
Losslessjoin decomposition:
R
1
R
2
= {A} and A ÷AB
Not dependency preserving
(cannot check B ÷C without computing R
1
R
2
)
©Silberschatz, Korth and Sudarshan 7.32 Database System Concepts
Testing for Dependency Preservation Testing for Dependency Preservation
To check if a dependency ¬÷¡ is preserved in a decomposition of R into R
1
, R
2
,
«, R
n
we apply the following simplified test (with attribute closure done w.r.t. F)
result = ¬
while (changes to result) do
for each R
i
in the decomposition
t = (result R
i
)
+
R
i
result = result t
If result contains all attributes in ¡, then the functional dependency
¬ ÷¡ is preserved.
We apply the test on all dependencies in F to check if a decomposition is
dependency preserving
This procedure takes polynomial time, instead of the exponential time required to
compute F
+
and (F
1
F
2
« F
n
)
+
©Silberschatz, Korth and Sudarshan 7.33 Database System Concepts
Boyce BoyceCodd Normal Form Codd Normal Form
¬ ÷¡ is trivial (i.e., ¡ _ ¬)
¬ is a superkey for R
A relation schema R is in BCNF with respect to a set F of functional
dependencies if for all functional dependencies in F
+
of the form
¬ ÷¡, where ¬ _ R and ¡ _ R, at least one of the following holds:
©Silberschatz, Korth and Sudarshan 7.34 Database System Concepts
Example Example
R = (A, B, C)
F = {A ÷B
B ÷C}
Key = {A}
R is not in BCNF
Decomposition R
1
= (A, B), R
2
= (B, C)
R
1
and R
2
in BCNF
Losslessjoin decomposition
Dependency preserving
©Silberschatz, Korth and Sudarshan 7.35 Database System Concepts
Testing for BCNF Testing for BCNF
To check if a nontrivial dependency ¬ ÷¡ causes a violation of BCNF
1. compute ¬
+
(the attribute closure of ¬), and
2. verify that it includes all attributes of R, that is, it is a superkey of R.
Simplified test: To check if a relation schema R with a given set of functional
dependencies F is in BCNF, it suffices to check only the dependencies in the given set F
for violation of BCNF, rather than checking all dependencies in F
+
.
We can show that if none of the dependencies in F causes a violation of BCNF, then none of
the dependencies in F
+
will cause a violation of BCNF either.
However, using only F is incorrect when testing a relation in a decomposition of R
E.g. Consider R (A, B, C, D), with F = { A ÷B, B ÷C}
Decompose R into R
1
(A,B) and R
2
(A,C,D)
Neither of the dependencies in F contain only attributes from (A,C,D) so we might be
mislead into thinking R
2
satisfies BCNF.
In fact, dependency A ÷C in F
+
shows R
2
is not in BCNF.
©Silberschatz, Korth and Sudarshan 7.36 Database System Concepts
BCNF Decomposition Algorithm BCNF Decomposition Algorithm
result := {R};
done := false;
compute F
+
;
while (not done) do
if (there is a schema R
i
in result that is not in BCNF)
then begin
let ¬ ÷¡ be a nontrivial functional
dependency that holds on R
i
such that ¬ ÷R
i
is not in F
+
,
and ¬ ¡ = ©;
result := (result ± R
i
) (R
i
± ¡) (¬, ¡ );
end
else done := true;
Note: each R
i
is in BCNF, and decomposition is losslessjoin.
©Silberschatz, Korth and Sudarshan 7.37 Database System Concepts
Example of BCNF Decomposition Example of BCNF Decomposition
R = (branchname, branchcity, assets,
customername, loannumber, amount)
F = {branchname ÷assets branchcity
loannumber ÷amount branchname}
Key = {loannumber, customername}
Decomposition
R
1
= (branchname, branchcity, assets)
R
2
= (branchname, customername, loannumber, amount)
R
3
= (branchname, loannumber, amount)
R
4
= (customername, loannumber)
Final decomposition
R
1
, R
3
, R
4
©Silberschatz, Korth and Sudarshan 7.38 Database System Concepts
Testing Decomposition for BCNF Testing Decomposition for BCNF
To check if a relation R
i
in a decomposition of R is in BCNF,
Either test R
i
for BCNF with respect to the restriction of F to R
i
(that is, all FDs in F
+
that contain only attributes from R
i
)
or use the original set of dependencies F that hold on R, but with the following test:
± for every set of attributes ¬ _ R
i
, check that ¬
+
(the attribute closure of ¬)
either includes no attribute of R
i
 ¬, or includes all attributes of R
i
.
If the condition is violated by some ¬ ÷¡ in F, the dependency
¬ ÷(¬
+
 ¬ ) R
i
can be shown to hold on R
i
, and R
i
violates BCNF.
We use above dependency to decompose R
i
©Silberschatz, Korth and Sudarshan 7.39 Database System Concepts
BCNF and Dependency Preservation BCNF and Dependency Preservation
R = (J, K, L)
F = {JK ÷L
L ÷K}
Two candidate keys = JK and JL
R is not in BCNF
Any decomposition of R will fail to preserve
JK ÷L
It is not always possible to get a BCNF decomposition that is
dependency preserving
©Silberschatz, Korth and Sudarshan 7.40 Database System Concepts
Third Normal Form: Motivation Third Normal Form: Motivation
There are some situations where
BCNF is not dependency preserving, and
efficient checking for FD violation on updates is important
Solution: define a weaker normal form, called Third Normal Form.
Allows some redundancy (with resultant problems; we will see examples later)
But FDs can be checked on individual relations without computing a join.
There is always a losslessjoin, dependencypreserving decomposition into 3NF.
©Silberschatz, Korth and Sudarshan 7.41 Database System Concepts
Third Normal Form Third Normal Form
A relation schema R is in third normal form (3NF) if for all:
¬ ÷¡ in F
+
at least one of the following holds:
¬ ÷¡ is trivial (i.e., ¡ ¬)
¬ is a superkey for R
Each attribute A in ¡ ± ¬ is contained in a candidate key for R.
(NOTE: each attribute may be in a different candidate key)
If a relation is in BCNF it is in 3NF (since in BCNF one of the first two
conditions above must hold).
Third condition is a minimal relaxation of BCNF to ensure dependency
preservation (will see why later).
©Silberschatz, Korth and Sudarshan 7.42 Database System Concepts
3NF (Cont.) 3NF (Cont.)
Example
R = (J, K, L)
F = {JK ÷L, L ÷K}
Two candidate keys: JK and JL
R is in 3NF
JK ÷L JK is a superkey
L ÷K K is contained in a candidate key
BCNF decomposition has (JL) and (LK)
Testing for JK ÷L requires a join
There is some redundancy in this schema
Equivalent to example in book:
Bankerschema = (branchname, customername, bankername)
bankername ÷branch name
branch name customername ÷bankername
©Silberschatz, Korth and Sudarshan 7.43 Database System Concepts
Testing for 3NF Testing for 3NF
Optimization: Need to check only FDs in F, need not check all FDs in F
+
.
Use attribute closure to check, for each dependency ¬ ÷¡, if ¬ is a superkey.
If ¬ is not a superkey, we have to verify if each attribute in ¡ is contained in a
candidate key of R
this test is rather more expensive, since it involve finding candidate keys
testing for 3NF has been shown to be NPhard
Interestingly, decomposition into third normal form (described shortly) can be done
in polynomial time
©Silberschatz, Korth and Sudarshan 7.44 Database System Concepts
3NF Decomposition Algorithm 3NF Decomposition Algorithm
Let F
c
be a canonical cover for F;
i := 0;
for each functional dependency ¬ ÷¡ in F
c
do
if none of the schemas R
j
, 1 e j e i contains ¬ ¡
then begin
i := i + 1;
R
i
:= ¬ ¡
end
if none of the schemas R
j
, 1 e j e i contains a candidate key for R
then begin
i := i + 1;
R
i
:= any candidate key for R;
end
return (R
1
, R
2
, ..., R
i
)
©Silberschatz, Korth and Sudarshan 7.45 Database System Concepts
3NF Decomposition Algorithm (Cont.) 3NF Decomposition Algorithm (Cont.)
Above algorithm ensures:
each relation schema R
i
is in 3NF
decomposition is dependency preserving and losslessjoin
Proof of correctness is at end of this file (click here)
©Silberschatz, Korth and Sudarshan 7.46 Database System Concepts
Example Example
Relation schema:
Bankerinfoschema = (branchname, customername,
bankername, officenumber)
The functional dependencies for this relation schema are:
bankername ÷branchname officenumber
customername branchname ÷bankername
The key is:
{customername, branchname}
©Silberschatz, Korth and Sudarshan 7.47 Database System Concepts
Applying 3NF to Applying 3NF to Banker Bankerinfo infoschema schema
The for loop in the algorithm causes us to include the following schemas
in our decomposition:
Bankerofficeschema = (bankername, branchname,
officenumber)
Bankerschema = (customername, branchname,
bankername)
Since Bankerschema contains a candidate key for
Bankerinfoschema, we are done with the decomposition process.
©Silberschatz, Korth and Sudarshan 7.48 Database System Concepts
Comparison of BCNF and 3NF Comparison of BCNF and 3NF
It is always possible to decompose a relation into relations in 3NF and
the decomposition is lossless
the dependencies are preserved
It is always possible to decompose a relation into relations in BCNF and
the decomposition is lossless
it may not be possible to preserve dependencies.
©Silberschatz, Korth and Sudarshan 7.49 Database System Concepts
Comparison of BCNF and 3NF (Cont.) Comparison of BCNF and 3NF (Cont.)
J
j
1
j
2
j
3
null
L
l
1
l
1
l
1
l
2
K
k
1
k
1
k
1
k
2
A schema that is in 3NF but not in BCNF has the problems of
repetition of information (e.g., the relationship l
1
, k
1
)
need to use null values (e.g., to represent the relationship
l
2
, k
2
where there is no corresponding value for J).
Example of problems due to redundancy in 3NF
R = (J, K, L)
F = {JK ÷L, L ÷K}
©Silberschatz, Korth and Sudarshan 7.50 Database System Concepts
Design Goals Design Goals
Goal for a relational database design is:
BCNF.
Lossless join.
Dependency preservation.
If we cannot achieve this, we accept one of
Lack of dependency preservation
Redundancy due to use of 3NF
Interestingly, SQL does not provide a direct way of specifying functional
dependencies other than superkeys.
Can specify FDs using assertions, but they are expensive to test
Even if we had a dependency preserving decomposition, using SQL we would
not be able to efficiently test a functional dependency whose left hand side is
not a key.
©Silberschatz, Korth and Sudarshan 7.51 Database System Concepts
Testing for FDs Across Relations Testing for FDs Across Relations
If decomposition is not dependency preserving, we can have an extra
materialized view for each dependency ¬ ÷¡ in F
c
that is not preserved in
the decomposition
The materialized view is defined as a projection on ¬ ¡ of the join of the
relations in the decomposition
Many newer database systems support materialized views and database
system maintains the view when the relations are updated.
No extra coding effort for programmer.
The FD becomes a candidate key on the materialized view.
Space overhead: for storing the materialized view
Time overhead: Need to keep materialized view up to date when relations are
updated
©Silberschatz, Korth and Sudarshan 7.52 Database System Concepts
Multivalued Dependencies Multivalued Dependencies
There are database schemas in BCNF that do not seem to be sufficiently
normalized
Consider a database
classes(course, teacher, book)
such that (c,t,b) classes means that t is qualified to teach c, and b is a
required textbook for c
The database is supposed to list for each course the set of teachers any one of
which can be the course¶s instructor, and the set of books, all of which are
required for the course (no matter who teaches it).
©Silberschatz, Korth and Sudarshan 7.53 Database System Concepts
Since there are nontrivial dependencies, (course, teacher, book) is the only
key, and therefore the relation is in BCNF
Insertion anomalies ± i.e., if Sara is a new teacher that can teach database, two
tuples need to be inserted
(database, Sara, DB Concepts)
(database, Sara, Ullman)
course teacher book
database
database
database
database
database
database
operating systems
operating systems
operating systems
operating systems
Avi
Avi
Hank
Hank
Sudarshan
Sudarshan
Avi
Avi
Jim
Jim
DB Concepts
Ullman
DB Concepts
Ullman
DB Concepts
Ullman
OS Concepts
Shaw
OS Concepts
Shaw
classes
©Silberschatz, Korth and Sudarshan 7.54 Database System Concepts
Therefore, it is better to decompose classes into:
course teacher
database
database
database
operating systems
operating systems
Avi
Hank
Sudarshan
Avi
Jim
teaches
course book
database
database
operating systems
operating systems
DB Concepts
Ullman
OS Concepts
Shaw
text
We shall see that these two relations are in Fourth Normal Form (4NF)
©Silberschatz, Korth and Sudarshan 7.55 Database System Concepts
Multivalued Dependencies (MVDs) Multivalued Dependencies (MVDs)
Let R be a relation schema and let ¬ _ R and ¡ _ R. The
multivalued dependency
¬ ÷÷¡
holds on R if in any legal relation r(R), for all pairs for tuples t
1
and t
2
in r such that t
1
[¬] = t
2
[¬], there exist tuples t
3
and t
4
in r such that:
t
1
[¬] = t
2
[¬] = t
3
[¬] t
4
[¬]
t
3
[¡] = t
1
[¡]
t
3
[R ± ¡] = t
2
[R ± ¡]
t
4
¡] = t
2
[¡]
t
4
[R ± ¡] = t
1
[R ± ¡]
©Silberschatz, Korth and Sudarshan 7.56 Database System Concepts
MVD (Cont.) MVD (Cont.)
Tabular representation of ¬ ÷÷¡
©Silberschatz, Korth and Sudarshan 7.57 Database System Concepts
Example Example
Let R be a relation schema with a set of attributes that are partitioned into 3
nonempty subsets.
Y, Z, W
We say that Y ÷÷Z (Y multidetermines Z)
if and only if for all possible relations r(R)
< y
1
, z
1
, w
1
> r and < y
2
, z
2
, w
2
> r
then
< y
1
, z
1
, w
2
> r and < y
1
, z
2
, w
1
> r
Note that since the behavior of Z and Ware identical it follows that Y ÷÷Z if Y
÷÷W
©Silberschatz, Korth and Sudarshan 7.58 Database System Concepts
Example (Cont.) Example (Cont.)
In our example:
course ÷÷teacher
course ÷÷book
The above formal definition is supposed to formalize the notion that
given a particular value of Y (course) it has associated with it a set of
values of Z (teacher) and a set of values of W (book), and these two
sets are in some sense independent of each other.
Note:
If Y ÷Z then Y ÷÷Z
Indeed we have (in above notation) Z
1
= Z
2
The claim follows.
©Silberschatz, Korth and Sudarshan 7.59 Database System Concepts
Use of Multivalued Dependencies Use of Multivalued Dependencies
We use multivalued dependencies in two ways:
1. To test relations to determine whether they are legal under a given set of
functional and multivalued dependencies
2. To specify constraints on the set of legal relations. We shall thus concern
ourselves only with relations that satisfy a given set of functional and
multivalued dependencies.
If a relation r fails to satisfy a given multivalued dependency, we can
construct a relations rd that does satisfy the multivalued dependency
by adding tuples to r.
©Silberschatz, Korth and Sudarshan 7.60 Database System Concepts
Theory of MVDs Theory of MVDs
From the definition of multivalued dependency, we can derive the following
rule:
If ¬ ÷¡, then ¬ ÷÷¡
That is, every functional dependency is also a multivalued dependency
The closure D
+
of D is the set of all functional and multivalued dependencies
logically implied by D.
We can compute D
+
from D, using the formal definitions of functional dependencies
and multivalued dependencies.
We can manage with such reasoning for very simple multivalued dependencies,
which seem to be most common in practice
For complex dependencies, it is better to reason about sets of dependencies
using a system of inference rules (see Appendix C).
©Silberschatz, Korth and Sudarshan 7.61 Database System Concepts
Fourth Normal Form Fourth Normal Form
A relation schema R is in 4NF with respect to a set D of functional and
multivalued dependencies if for all multivalued dependencies in D
+
of the form
¬ ÷÷¡, where ¬ _ R and ¡ _ R, at least one of the following hold:
¬ ÷÷ ¡ is trivial (i.e., ¡ _ ¬ or ¬ ¡ = R)
¬ is a superkey for schema R
If a relation is in 4NF it is in BCNF
©Silberschatz, Korth and Sudarshan 7.62 Database System Concepts
Restriction of Multivalued Dependencies Restriction of Multivalued Dependencies
The restriction of D to R
i
is the set D
i
consisting of
All functional dependencies in D
+
that include only attributes of R
i
All multivalued dependencies of the form
¬ ÷÷ (¡ R
i
)
where ¬ _ R
i
and ¬ ÷÷ ¡ is in D
+
©Silberschatz, Korth and Sudarshan 7.63 Database System Concepts
4NF Decomposition Algorithm 4NF Decomposition Algorithm
result: = {R};
done := false;
compute D
+
;
Let D
i
denote the restriction of D
+
to R
i
while (not done)
if (there is a schema R
i
in result that is not in 4NF) then
begin
let ¬ ÷÷¡ be a nontrivial multivalued dependency that holds
on R
i
such that ¬ ÷R
i
is not in D
i
, and ¬¡=o;
result := (result  R
i
) (R
i
 ¡) (¬, ¡);
end
else done:= true;
Note: each R
i
is in 4NF, and decomposition is losslessjoin
©Silberschatz, Korth and Sudarshan 7.64 Database System Concepts
Example Example
R =(A, B, C, G, H, I)
F ={ A ÷÷B
B ÷÷HI
CG ÷÷H }
R is not in 4NF since A ÷÷ B and A is not a superkey for R
Decomposition
a) R
1
= (A, B) (R
1
is in 4NF)
b) R
2
= (A, C, G, H, I) (R
2
is not in 4NF)
c) R
3
= (C, G, H) (R
3
is in 4NF)
d) R
4
= (A, C, G, I) (R
4
is not in 4NF)
Since A ÷÷ B and B ÷÷ HI, A ÷÷ HI, A ÷÷ I
e) R
5
= (A, I) (R
5
is in 4NF)
f)R
6
= (A, C, G) (R
6
is in 4NF)
©Silberschatz, Korth and Sudarshan 7.65 Database System Concepts
Further Normal Forms Further Normal Forms
join dependencies generalize multivalued dependencies
lead to projectjoin normal form (PJNF) (also called fifth normal form)
A class of even more general constraints, leads to a normal form called
domainkey normal form.
Problem with these generalized constraints: i hard to reason with, and no set
of sound and complete set of inference rules.
Hence rarely used
©Silberschatz, Korth and Sudarshan 7.66 Database System Concepts
Overall Database Design Process Overall Database Design Process
We have assumed schema R is given
R could have been generated when converting ER diagram to a set of tables.
R could have been a single relation containing all attributes that are of interest (called
universal relation).
Normalization breaks R into smaller relations.
R could have been the result of some ad hoc design of relations, which we then
test/convert to normal form.
©Silberschatz, Korth and Sudarshan 7.67 Database System Concepts
ER Model and Normalization ER Model and Normalization
When an ER diagram is carefully designed, identifying all entities correctly, the tables
generated from the ER diagram should not need further normalization.
However, in a real (imperfect) design there can be FDs from nonkey attributes of an
entity to other attributes of the entity
E.g. employee entity with attributes departmentnumber and departmentaddress,
and an FD departmentnumber ÷departmentaddress
Good design would have made department an entity
FDs from nonkey attributes of a relationship set possible, but rare  most
relationships are binary
©Silberschatz, Korth and Sudarshan 7.68 Database System Concepts
Universal Relation Approach Universal Relation Approach
Dangling tuples ± Tuples that ³disappear´ in computing a join.
Let r
1
(R
1
), r
2
(R
2
), «., r
n
(R
n
) be a set of relations
A tuple r of the relation r
i
is a dangling tuple if r is not in the relation:

Ri
(r
1
r
2
« r
n
)
The relation r
1
r
2
« r
n
is called a universal relation since it involves all the
attributes in the ³universe´ defined by
R
1
R
2
« R
n
If dangling tuples are allowed in the database, instead of decomposing a
universal relation, we may prefer to synthesize a collection of normal form
schemas from a given set of attributes.
©Silberschatz, Korth and Sudarshan 7.69 Database System Concepts
Universal Relation Approach Universal Relation Approach
Dangling tuples may occur in practical database applications.
They represent incomplete information
E.g. may want to break up information about loans into:
(branchname, loannumber)
(loannumber, amount)
(loannumber, customername)
Universal relation would require null values, and have dangling tuples
©Silberschatz, Korth and Sudarshan 7.70 Database System Concepts
Universal Relation Approach (Contd.) Universal Relation Approach (Contd.)
A particular decomposition defines a restricted form of incomplete information
that is acceptable in our database.
Above decomposition requires at least one of customername, branchname or
amount in order to enter a loan number without using null values
Rules out storing of customername, amount without an appropriate loannumber
(since it is a key, it can't be null either!)
Universal relation requires unique attribute names unique role assumption
e.g. customername, branchname
Reuse of attribute names is natural in SQL since relation names can be
prefixed to disambiguate names
©Silberschatz, Korth and Sudarshan 7.71 Database System Concepts
Denormalization for Performance Denormalization for Performance
May want to use nonnormalized schema for performance
E.g. displaying customername along with accountnumber and balance requires
join of account with depositor
Alternative 1: Use denormalized relation containing attributes of account as well
as depositor with all above attributes
faster lookup
Extra space and extra execution time for updates
extra coding work for programmer and possibility of error in extra code
Alternative 2: use a materialized view defined as
account depositor
Benefits and drawbacks same as above, except no extra coding work for
programmer and avoids possible errors
©Silberschatz, Korth and Sudarshan 7.72 Database System Concepts
Other Design Issues Other Design Issues
Some aspects of database design are not caught by normalization
Examples of bad database design, to be avoided:
Instead of earnings(companyid, year, amount), use
earnings2000, earnings2001, earnings2002, etc., all on the schema (company
id, earnings).
Above are in BCNF, but make querying across years difficult and needs new
table each year
companyyear(companyid, earnings2000, earnings2001, earnings2002)
Also in BCNF, but also makes querying across years difficult and requires new
attribute each year.
Is an example of a crosstab, where values for one attribute become column
names
Used in spreadsheets, and in data analysis tools
Proof of Correctness of 3NF Decomposition Proof of Correctness of 3NF Decomposition
Algorithm Algorithm
©Silberschatz, Korth and Sudarshan 7.74 Database System Concepts
Correctness of 3NF Decomposition Algorithm Correctness of 3NF Decomposition Algorithm
3NF decomposition algorithm is dependency preserving (since there is a
relation for every FD in F
c
)
Decomposition is lossless join
A candidate key (C) is in one of the relations R
i
in decomposition
Closure of candidate key under F
c
must contain all attributes in R.
Follow the steps of attribute closure algorithm to show there is only one tuple in the
join result for each tuple in R
i
©Silberschatz, Korth and Sudarshan 7.75 Database System Concepts
Correctness of 3NF Decomposition Algorithm Correctness of 3NF Decomposition Algorithm
(Contd.) (Contd.)
Claim: if a relation R
i
is in the decomposition generated by the
above algorithm, then R
i
satisfies 3NF.
Let R
i
be generated from the dependency ¬ ÷¡
Let K ÷¡ be any nontrivial functional dependency on R
i
. (We need only
consider FDs whose righthand side is a single attribute.)
Now, B can be in either ¡ or ¬ but not in both. Consider each case separately.
©Silberschatz, Korth and Sudarshan 7.76 Database System Concepts
Correctness of 3NF Decomposition (Contd.) Correctness of 3NF Decomposition (Contd.)
Case 1: If B in ¡:
If K is a superkey, the 2nd condition of 3NF is satisfied
Otherwise ¬ must contain some attribute not in K
Since K ÷B is in F
+
it must be derivable from F
c
, by using attribute closure on K.
Attribute closure not have used ¬ ÷¡  if it had been used, ¬ must be contained
in the attribute closure of K, which is not possible, since we assumed K is not a
superkey.
Now, using ¬÷ (¡ {B}) and K ÷B, we can derive ¬ ÷B
(since K _ ¬ ¡, and ¡ Z K since K ÷B is nontrivial)
Then, B is extraneous in the righthand side of ¬ ÷¡; which is not possible since
¬ ÷¡ is in F
c
.
Thus, if B is in ¡ then K must be a superkey, and the second condition of 3NF
must be satisfied.
©Silberschatz, Korth and Sudarshan 7.77 Database System Concepts
Correctness of 3NF Decomposition (Contd.) Correctness of 3NF Decomposition (Contd.)
Case 2: B is in ¬.
Since ¬ is a candidate key, the third alternative in the definition of 3NF is trivially
satisfied.
In fact, we cannot show that K is a superkey.
This shows exactly why the third alternative is present in the definition of 3NF.
Q.E.D.
End of Chapter End of Chapter
©Silberschatz, Korth and Sudarshan 7.79 Database System Concepts
Sample Sample lending lending Relation Relation
©Silberschatz, Korth and Sudarshan 7.80 Database System Concepts
Sample Relation Sample Relation rr
©Silberschatz, Korth and Sudarshan 7.81 Database System Concepts
The The customer customer Relation Relation
©Silberschatz, Korth and Sudarshan 7.82 Database System Concepts
The The loan loan Relation Relation
©Silberschatz, Korth and Sudarshan 7.83 Database System Concepts
The The branch branch Relation Relation
©Silberschatz, Korth and Sudarshan 7.84 Database System Concepts
The Relation The Relation branch branchcustomer customer
©Silberschatz, Korth and Sudarshan 7.85 Database System Concepts
The Relation The Relation customer customerloan loan
©Silberschatz, Korth and Sudarshan 7.86 Database System Concepts
The Relation The Relation branch branchcustomer customer customer customerloan loan
©Silberschatz, Korth and Sudarshan 7.87 Database System Concepts
An Instance of An Instance of Banker Bankerschema schema
©Silberschatz, Korth and Sudarshan 7.88 Database System Concepts
Tabular Representation of Tabular Representation of ¬ ÷÷ ¡ ¬ ÷÷ ¡
©Silberschatz, Korth and Sudarshan 7.89 Database System Concepts
Relation Relation bc bc: An Example of Reduncy in a BCNF Relation : An Example of Reduncy in a BCNF Relation
©Silberschatz, Korth and Sudarshan 7.90 Database System Concepts
An Illegal An Illegal bc bc Relation Relation
©Silberschatz, Korth and Sudarshan 7.91 Database System Concepts
Decomposition of Decomposition of loan loaninfo info
©Silberschatz, Korth and Sudarshan 7.92 Database System Concepts
Relation of Exercise 7.4 Relation of Exercise 7.4
Chapter 7: Relational Database Design
First Normal Form Pitfalls in Relational Database Design Functional Dependencies Decomposition BoyceCodd Normal Form Third Normal Form Multivalued Dependencies and Fourth Normal Form Overall Database Design Process
Database System Concepts
7.2
©Silberschatz, Korth and Sudarshan
First Normal Form
Domain is atomic if its elements are considered to be indivisible units Examples of nonatomic domains: Set of names, composite attributes Identification numbers like CS101 that can be broken up into parts A relational schema R is in first normal form if the domains of all attributes of R
are atomic
Nonatomic values complicate storage and encourage redundant (repeated)
storage of data
E.g. Set of accounts stored with each customer, and set of owners stored with each account We assume all relations are in first normal form (revisit this in Chapter 9 on Object Relational Databases)
Database System Concepts
7.3
©Silberschatz, Korth and Sudarshan
First Normal Form (Contd.4 ©Silberschatz.g. Doing so is a bad idea: leads to encoding of information in application program rather than in the database. Database System Concepts 7.) Atomicity is actually a property of how the elements of the domain are used. E. the domain of roll numbers is not atomic. Strings would normally be considered indivisible Suppose that students are given roll numbers which are strings CS0012 or EE1127 of the form If the first two characters are extracted to find the department. Korth and Sudarshan .
Database System Concepts 7.5 ©Silberschatz. Design Goals: Avoid redundant data Ensure that relationships among attributes are represented Facilitate the checking of updates for violation of database integrity constraints.Pitfalls in Relational Database Design Relational database design requires that we find a ³good´ collection of relation schemas. Korth and Sudarshan . Inability to represent certain information. A bad design may lead to Repetition of Information.
Example Consider the relation schema: Lendingschema = (branchname. assets. branchcity. Database System Concepts 7. customername. amount) Redundancy: Data for branchname. loannumber.6 ©Silberschatz. introducing possibility of inconsistency of assets value Null values Cannot store information about a branch if no loans exist Can use null values. Korth and Sudarshan . branchcity. assets are repeated for each loan that a branch makes Wastes space Complicates updating. but they are difficult to handle.
For all possible relations r on schema R r = R1 (r) R2 (r) Database System Concepts 7. branchname. branchcity.7 ©Silberschatz. R2): R = R1 R2 Losslessjoin decomposition. Korth and Sudarshan . amount) All attributes of an original schema (R) must appear in the decomposition (R1. loannumber.assets) Loaninfoschema = (customername.Decomposition Decompose the relation schema Lendingschema into: Branchschema = (branchname.
8 ©Silberschatz. Korth and Sudarshan .Example of Non LosslessJoin Decomposition Lossless Decomposition of R = (A. B) R2 = (A) A E E F r B 1 2 1 A E F A(r) A E E F F B 1 2 1 2 R2 = (B) B 1 2 B(r) A (r) B (r) Database System Concepts 7.
Goal ³ Devise a Theory for the Following Decide whether a particular relation R is in ³good´ form. .. In the case that a relation R is not in ³good´ form.. decompose it into a set of relations {R1.. R2. Korth and Sudarshan . Rn} such that each relation is in good form the decomposition is a losslessjoin decomposition Our theory is based on: functional dependencies multivalued dependencies Database System Concepts 7.9 ©Silberschatz.
10 ©Silberschatz. Korth and Sudarshan . Database System Concepts 7. A functional dependency is a generalization of the notion of a key. Require that the value for a certain set of attributes determines uniquely the value for another set of attributes.Functional Dependencies Constraints on the set of legal relations.
) Let R be a relation schema E R and F R The functional dependency EpF holds on R if and only if for any legal relations r(R). That is. A p B does NOT hold. t1[E] = t2 [E] t1[F ] = t2 [F ] Example: Consider r(A.B) with the following instance of r. Korth and Sudarshan . they also agree on the attributes F. whenever any two tuples t1 and t2 of r agree on the attributes E.11 ©Silberschatz. but B p A does hold. Database System Concepts 7.Functional Dependencies (Cont. 1 4 1 5 3 7 On this instance.
Korth and Sudarshan . We expect this set of functional dependencies to hold: loannumber p amount loannumber p branchname but would not expect the following to hold: loannumber p customername Database System Concepts 7.Functional Dependencies (Cont.12 ©Silberschatz. loannumber. Consider the schema: Loaninfoschema = (customername.) K is a superkey for relation schema R if and only if K p R K is a candidate key for R if and only if K p R. amount). E p R Functional dependencies allow us to express constraints that cannot be expressed using superkeys. and for no E K. branchname.
13 ©Silberschatz. For example. Korth and Sudarshan . satisfy loannumber p customername. Database System Concepts 7.Use of Functional Dependencies We use functional dependencies to: test relations to see if they are legal under a given set of functional dependencies. Note: A specific instance of a relation schema may satisfy a functional dependency even if the functional dependency does not hold on all legal instances. by chance. a specific instance of Loanschema may. we say that r satisfies F. If a relation r is legal under a set F of functional dependencies. specify constraints on the set of legal relations We say that F holds on R if all legal relations on R satisfy the set of functional dependencies F.
E p F is trivial if F E Database System Concepts 7.) A functional dependency is trivial if it is satisfied by all instances of a relation E.g. loannumber p customername customername p customername In general.14 ©Silberschatz. customername.Functional Dependencies (Cont. Korth and Sudarshan .
and F p K. (reflexivity) (augmentation) if E p F. then we can infer that A p C The set of all functional dependencies logically implied by F is the closure of F. Korth and Sudarshan . We can find all of F+ by applying Armstrong¶s Axioms: if F E. then K E p K F These rules are sound (generate only functional dependencies that actually hold) and complete (generate all functional dependencies that hold). E. then E p F if E p F. then E p K (transitivity) Database System Concepts 7. If A p B and B p C.g. We denote the closure of F by F+.Closure of a Set of Functional Dependencies Given a set F set of functional dependencies. there are certain other functional dependencies that are logically implied by F.15 ©Silberschatz.
H. Korth and Sudarshan . B. C. or ± Augmentation of CG p I to infer CG p CGI. augmentation of CG p H to infer CGI p HI. I) F={ ApB ApC CG p H CG p I B p H} some members of F+ ApH by transitivity from A p B and B p H AG p I by augmenting A p C with G. to get AG p CG and then transitivity with CG p I CG p HI from CG p H and CG p I : ³union rule´ can be inferred from ± definition of functional dependencies. G. and then transitivity Database System Concepts 7.16 ©Silberschatz.Example R = (A.
Procedure for Computing F+ To compute the closure of a set of functional dependencies F: F+ = F repeat for each functional dependency f in F+ apply reflexivity and augmentation rules on f add the resulting functional dependencies to F+ for each pair of functional dependencies f1and f2 in F+ if f1 and f2 can be combined using transitivity then add the resulting functional dependency to F+ until F+ does not change any further NOTE: We will see an alternative procedure for this task later Database System Concepts 7. Korth and Sudarshan .17 ©Silberschatz.
then E p F K holds (union) If E p F K holds. Korth and Sudarshan . Database System Concepts 7.Closure of Functional Dependencies (Cont. then E K p H holds (pseudotransitivity) The above rules can be inferred from Armstrong¶s axioms.18 ©Silberschatz. If E p F holds and E p K holds.) We can further simplify manual computation of F+ by using the following additional rules. then E p F holds and E p K holds (decomposition) If E p F holds and K F p H holds.
Closure of Attribute Sets Given a set of attributes E define the closure of E under F (denoted by E+) as the set of attributes that are functionally determined by E under F: E p F is in F+ F E+ Algorithm to compute E+. the closure of E under F result := E.19 ©Silberschatz. while (changes to result) do for each F p K in F do begin if F result then result := result K end Database System Concepts 7. Korth and Sudarshan .
20 ©Silberschatz. Is AG a super key? Does AG p R? 2. result = ABCGHI (A p C and A p B) (CG p H and CG AGBC) (CG p I and CG AGBCH) Is AG a candidate key? 1. Database System Concepts 7. Does G+ p R? 1. B. G. result = ABCG 3. result = ABCGH 4. C. Does A+ p R? 2. Korth and Sudarshan .Example of Attribute Set Closure R = (A. Is any subset of AG a superkey? 1. result = AG 2. H. I) F = {A p B ApC CG p H CG p I B p H} (AG)+ 1.
and very useful Computing closure of F For each K R. we find the closure K+. just check if F E+. Testing functional dependencies To check if a functional dependency E p F holds (or. we compute E+ by using attribute closure. and check if E+ contains all attributes of R. we compute E+. Database System Concepts 7. we output a functional dependency K p S.21 ©Silberschatz. in other words. Is a simple and cheap test. and then check if it contains F. That is. and for each S K+.Uses of Attribute Closure There are several uses of the attribute closure algorithm: Testing for superkey: To test if E is a superkey. Korth and Sudarshan . is in F+).
22 ©Silberschatz. AC p D} can be simplified to {A p B. B p C. A p D} Intuitively. A p D} E. Korth and Sudarshan . a canonical cover of F is a ³minimal´ set of functional dependencies equivalent to F.g. A p C} Parts of a functional dependency may be redundant E. B p C.Canonical Cover Sets of functional dependencies may have redundant dependencies that can be inferred from the others Eg: A p C is redundant in: {A p B. B p C. A p CD} can be simplified to {A p B. with no redundant dependencies or having redundant parts of dependencies Database System Concepts 7. on RHS: {A p B. B p C.g. on LHS: {A p B. B p C.
since a ³stronger´ functional dependency always implies a weaker one Example: Given F = {A p C. Example: Given F = {A p C. AB p CD} C is extraneous in AB p CD since A p C can be inferred even after deleting C Database System Concepts 7. Attribute A is extraneous in F if A F and the set of functional dependencies (F ± {E p F}) {E p(F ± A)} logically implies F. Korth and Sudarshan .Extraneous Attributes Consider a set F of functional dependencies and the functional dependency E p F in F. Attribute A is extraneous in E if A E and F logically implies (F ± {E p F}) {(E ± A) p F}.23 ©Silberschatz. AB p C } B is extraneous in AB p C because A p C logically implies AB p C. Note: implication in the opposite direction is trivial in each of the cases above.
compute (A ± {E})+ using the dependencies in F check that (A ± {E})+ contains E.Testing if an Attribute is Extraneous Consider a set F of functional dependencies and the functional dependency E p F in F. To test if attribute A E is extraneous in E 1. Database System Concepts 7. A is extraneous compute E+ using only the dependencies in F¶ = (F ± {E p F}) {E p(F ± A)}. if it does. check that E+ contains A. if it does. A is extraneous To test if attribute A F is extraneous in F 1. 2. Korth and Sudarshan .24 ©Silberschatz. 2.
and No functional dependency in Fc contains an extraneous attribute. Korth and Sudarshan . delete it from E p F until F does not change Note: Union rule may become applicable after some extraneous attributes have been deleted.Canonical Cover A canonical cover for F is a set of dependencies Fc such that F logically implies all dependencies in Fc. and Fc logically implies all dependencies in F. so it has to be reapplied Database System Concepts 7.25 ©Silberschatz. To compute a canonical cover for F: repeat Use the union rule to replace any dependencies in F E1 p F1 and E1 p F1 with E1 p F1 F2 Find a functional dependency E p F with an extraneous attribute either in E or in F If an extraneous attribute is found. and Each left side of functional dependency in Fc is unique.
AB p C} A is extraneous in AB p C because B p C logically implies AB p C. C) F = {A p BC BpC ApB AB p C} Combine A p BC and A p B into A p BC Set is now {A p BC. Korth and Sudarshan . B p C.Example of Computing a Canonical Cover R = (A. Set is now {A p BC. The canonical cover is: ApB BpC Database System Concepts 7.26 ©Silberschatz. B p C} C is extraneous in A p BC since A p BC is logically implied by A p B and B p C. B.
In the case that a relation R is not in ³good´ form. Korth and Sudarshan . R2... .. Rn} such that each relation is in good form the decomposition is a losslessjoin decomposition Our theory is based on: functional dependencies multivalued dependencies Database System Concepts 7.Goals of Normalization Decide whether a particular relation R is in ³good´ form. decompose it into a set of relations {R1.27 ©Silberschatz.
For all possible relations r on schema R r = R1 (r) R2 (r) A decomposition of R into R1 and R2 is lossless join if and only if at least one of the following dependencies is in F+: R1 R2 p R1 R1 R2 p R2 Database System Concepts 7. Korth and Sudarshan . amount) All attributes of an original schema (R) must appear in the decomposition (R1.28 ©Silberschatz. branchcity. R2): R = R1 R2 Losslessjoin decomposition.Decomposition Decompose the relation schema Lendingschema into: Branchschema = (branchname. branchname. loannumber.assets) Loaninfoschema = (customername.
Example: Decomposition of R = (A.Example of LossyJoin Decomposition Lossy Lossyjoin decompositions result in information loss.29 ©Silberschatz. B) R2 = (A) A E E F r B 1 2 1 A E F A(r) A E E F F B 1 2 1 2 R2 = (B) B 1 2 B(r) A (r) B (r) Database System Concepts 7. Korth and Sudarshan .
Database System Concepts 7. No redundancy: The relations Ri preferably should be in either BoyceCodd Normal Form or Third Normal Form. checking updates for violation of functional dependencies may require computing joins.Normalization Using Functional Dependencies When we decompose a relation schema R with a set of functional dependencies F into R1.. Dependency preservation: Let Fi be the set of dependencies F+ that include only attributes in Ri.. Korth and Sudarshan . which is expensive. that is. Preferably the decomposition should be dependency preserving. R2. Rn we want Losslessjoin decomposition: Otherwise decomposition would result in information loss..30 ©Silberschatz. F2 « Fn)+ = F+ (F1 Otherwise.
B). C) Losslessjoin decomposition: R1 R2 = {B} and B p BC Dependency preserving R1 = (A. B). B. Korth and Sudarshan .31 ©Silberschatz. C) Losslessjoin decomposition: R1 R2 = {A} and A p AB Not dependency preserving (cannot check B p C without computing R1 R2) Database System Concepts 7.Example R = (A. R2 = (B. C) F = {A p B. B p C) R1 = (A. R2 = (A.
We apply the test on all dependencies in F to check if a decomposition is dependency preserving This procedure takes polynomial time. «. Korth and Sudarshan . R2. instead of the exponential time required to compute F+ and (F1 F2 « Fn)+ Database System Concepts 7. Rn we apply the following simplified test (with attribute closure done w.r.32 ©Silberschatz. F) result = E while (changes to result) do for each Ri in the decomposition t = (result Ri)+ Ri result = result t If result contains all attributes in F.t.Testing for Dependency Preservation To check if a dependency EpF is preserved in a decomposition of R into R1. then the functional dependency E p F is preserved.
at least one of the following holds: E p F is trivial (i. F E) E is a superkey for R Database System Concepts 7..33 ©Silberschatz. Korth and Sudarshan .e. where E R and F R.BoyceBoyceCodd Normal Form A relation schema R is in BCNF with respect to a set F of functional dependencies if for all functional dependencies in F+ of the form Ep F.
C) R1 and R2 in BCNF Losslessjoin decomposition Dependency preserving Database System Concepts 7. C) F = {A p B B p C} Key = {A} R is not in BCNF Decomposition R1 = (A.Example R = (A. Korth and Sudarshan . R2 = (B. B). B.34 ©Silberschatz.
using only F is incorrect when testing a relation in a decomposition of R E. B pC} Decompose R into R1(A. C. However.C. Korth and Sudarshan . rather than checking all dependencies in F+. Consider R (A.D) Neither of the dependencies in F contain only attributes from (A.Testing for BCNF To check if a nontrivial dependency EpF causes a violation of BCNF 1. In fact. We can show that if none of the dependencies in F causes a violation of BCNF. Database System Concepts 7. that is.g. B. dependency A p C in F+ shows R2 is not in BCNF. D).35 ©Silberschatz.D) so we might be mislead into thinking R2 satisfies BCNF. with F = { A pB. verify that it includes all attributes of R. it is a superkey of R. and 2. then none of the dependencies in F+ will cause a violation of BCNF either. compute E+ (the attribute closure of E). it suffices to check only the dependencies in the given set F for violation of BCNF.C. Simplified test: To check if a relation schema R with a given set of functional dependencies F is in BCNF.B) and R2(A.
compute F+. done := false. and decomposition is losslessjoin. Note: each Ri is in BCNF. F ). and E F = .36 ©Silberschatz. result := (result ± Ri) (Ri ± F) (E. while (not done) do if (there is a schema Ri in result that is not in BCNF) then begin let E p F be a nontrivial functional dependency that holds on Ri such that E p Ri is not in F+. end else done := true. Korth and Sudarshan .BCNF Decomposition Algorithm result := {R}. Database System Concepts 7.
loannumber. customername. assets. loannumber. amount) F = {branchname p assets branchcity loannumber p amount branchname} Key = {loannumber. customername. assets) R2 = (branchname. R4 Database System Concepts 7. branchcity. loannumber) Final decomposition R1.Example of BCNF Decomposition R = (branchname. customername} Decomposition R1 = (branchname. amount) R4 = (customername. loannumber. amount) R3 = (branchname. Korth and Sudarshan .37 ©Silberschatz. R3. branchcity.
all FDs in F+ that contain only attributes from Ri) or use the original set of dependencies F that hold on R. Either test Ri for BCNF with respect to the restriction of F to Ri (that is. or includes all attributes of Ri.38 ©Silberschatz.Testing Decomposition for BCNF To check if a relation Ri in a decomposition of R is in BCNF. check that E+ (the attribute closure of E) either includes no attribute of Ri. We use above dependency to decompose Ri Database System Concepts 7. the dependency Ep (E+ . but with the following test: ± for every set of attributes E Ri.E. and Ri violates BCNF.E) Ri can be shown to hold on Ri. Korth and Sudarshan . If the condition is violated by some Ep F in F.
Korth and Sudarshan .39 ©Silberschatz. K.BCNF and Dependency Preservation It is not always possible to get a BCNF decomposition that is dependency preserving R = (J. L) F = {JK p L L p K} Two candidate keys = JK and JL R is not in BCNF Any decomposition of R will fail to preserve JK p L Database System Concepts 7.
we will see examples later) But FDs can be checked on individual relations without computing a join. Database System Concepts 7.Third Normal Form: Motivation There are some situations where BCNF is not dependency preserving.40 ©Silberschatz. Korth and Sudarshan . There is always a losslessjoin. and efficient checking for FD violation on updates is important Solution: define a weaker normal form. dependencypreserving decomposition into 3NF. Allows some redundancy (with resultant problems. called Third Normal Form.
.Third Normal Form A relation schema R is in third normal form (3NF) if for all: E p F in F+ at least one of the following holds: E p F is trivial (i. Korth and Sudarshan . Third condition is a minimal relaxation of BCNF to ensure dependency preservation (will see why later). Database System Concepts 7. (NOTE: each attribute may be in a different candidate key) If a relation is in BCNF it is in 3NF (since in BCNF one of the first two conditions above must hold).41 ©Silberschatz.e. F E) E is a superkey for R Each attribute A in F ± E is contained in a candidate key for R.
) Example R = (J.42 ©Silberschatz. L) F = {JK p L. customername.3NF (Cont. bankername) bankername p branch name branch name customername p bankername Database System Concepts 7. L p K} Two candidate keys: JK and JL R is in 3NF JK p L LpK JK is a superkey K is contained in a candidate key BCNF decomposition has (JL) and (LK) Testing for JK p L requires a join There is some redundancy in this schema Equivalent to example in book: Bankerschema = (branchname. K. Korth and Sudarshan .
for each dependency E p F. Korth and Sudarshan . we have to verify if each attribute in F is contained in a candidate key of R this test is rather more expensive. decomposition into third normal form (described shortly) can be done in polynomial time Database System Concepts 7. Use attribute closure to check.43 ©Silberschatz. since it involve finding candidate keys testing for 3NF has been shown to be NPhard Interestingly. need not check all FDs in F+. if E is a superkey.Testing for 3NF Optimization: Need to check only FDs in F. If E is not a superkey.
end return (R1. Ri := any candidate key for R.. 1 e j e i contains E F then begin i := i + 1. Ri := E F end if none of the schemas Rj. .44 ©Silberschatz. R2..3NF Decomposition Algorithm Let Fc be a canonical cover for F. Ri) Database System Concepts 7. i := 0. 1 e j e i contains a candidate key for R then begin i := i + 1. for each functional dependency E p F in Fc do if none of the schemas Rj. Korth and Sudarshan ..
) Above algorithm ensures: each relation schema Ri is in 3NF decomposition is dependency preserving and losslessjoin Proof of correctness is at end of this file (click here) Database System Concepts 7.3NF Decomposition Algorithm (Cont. Korth and Sudarshan .45 ©Silberschatz.
Example
Relation schema: Bankerinfoschema = (branchname, customername, bankername, officenumber) The functional dependencies for this relation schema are:
bankername p branchname officenumber customername branchname p bankername
The key is:
{customername, branchname}
Database System Concepts
7.46
©Silberschatz, Korth and Sudarshan
Applying 3NF to Bankerinfoschema Bankerinfo The for loop in the algorithm causes us to include the following schemas
in our decomposition: Bankerofficeschema = (bankername, branchname, officenumber) Bankerschema = (customername, branchname, bankername)
Since Bankerschema contains a candidate key for
Bankerinfoschema, we are done with the decomposition process.
Database System Concepts
7.47
©Silberschatz, Korth and Sudarshan
Comparison of BCNF and 3NF
It is always possible to decompose a relation into relations in 3NF and the decomposition is lossless the dependencies are preserved It is always possible to decompose a relation into relations in BCNF and the decomposition is lossless it may not be possible to preserve dependencies.
Database System Concepts
7.48
©Silberschatz, Korth and Sudarshan
. k2 where there is no corresponding value for J). Korth and Sudarshan . K. L p K} J j1 j2 j3 null L l1 l1 l1 l2 K k1 k1 k1 k2 A schema that is in 3NF but not in BCNF has the problems of repetition of information (e.49 ©Silberschatz. L) F = {JK p L. Database System Concepts 7.) Example of problems due to redundancy in 3NF R = (J. the relationship l1. to represent the relationship l2.Comparison of BCNF and 3NF (Cont..g. k1) need to use null values (e.g.
we accept one of Lack of dependency preservation Redundancy due to use of 3NF Interestingly.50 ©Silberschatz. Lossless join. using SQL we would not be able to efficiently test a functional dependency whose left hand side is not a key. Can specify FDs using assertions. Dependency preservation.Design Goals Goal for a relational database design is: BCNF. Database System Concepts 7. If we cannot achieve this. but they are expensive to test Even if we had a dependency preserving decomposition. SQL does not provide a direct way of specifying functional dependencies other than superkeys. Korth and Sudarshan .
51 ©Silberschatz. Korth and Sudarshan . No extra coding effort for programmer. we can have an extra materialized view for each dependency E pF in Fc that is not preserved in the decomposition The materialized view is defined as a projection on E F of the join of the relations in the decomposition Many newer database systems support materialized views and database system maintains the view when the relations are updated. Space overhead: for storing the materialized view Time overhead: Need to keep materialized view up to date when relations are updated Database System Concepts 7. The FD becomes a candidate key on the materialized view.Testing for FDs Across Relations If decomposition is not dependency preserving.
and the set of books.52 ©Silberschatz.t.Multivalued Dependencies There are database schemas in BCNF that do not seem to be sufficiently normalized Consider a database classes(course. book) such that (c. and b is a required textbook for c The database is supposed to list for each course the set of teachers any one of which can be the course¶s instructor. teacher. all of which are required for the course (no matter who teaches it). Database System Concepts 7.b) classes means that t is qualified to teach c. Korth and Sudarshan .
(course. teacher.course database database database database database database operating systems operating systems operating systems operating systems teacher Avi Avi Hank Hank Sudarshan Sudarshan Avi Avi Jim Jim classes book DB Concepts Ullman DB Concepts Ullman DB Concepts Ullman OS Concepts Shaw OS Concepts Shaw Since there are nontrivial dependencies. Korth and Sudarshan .e.. Sara. DB Concepts) (database. Sara. two tuples need to be inserted (database. and therefore the relation is in BCNF Insertion anomalies ± i. Ullman) Database System Concepts 7.53 ©Silberschatz. if Sara is a new teacher that can teach database. book) is the only key.
it is better to decompose classes into: course database database database operating systems operating systems course database database operating systems operating systems text teacher Avi Hank Sudarshan Avi Jim teaches book DB Concepts Ullman OS Concepts Shaw We shall see that these two relations are in Fourth Normal Form (4NF) Database System Concepts 7.54 ©Silberschatz. Korth and Sudarshan . Therefore.
Korth and Sudarshan . for all pairs for tuples t1 and t2 in r such that t1[E] = t2 [E].55 ©Silberschatz. there exist tuples t3 and t4 in r such that: t1[E] = t2 [E] = t3 [E] t4 [E] t3[F] = t1 [F] t3[R ± F] = t2[R ± F] t4 F] = t2[F] t4[R ± F] = t1[R ± F] Database System Concepts 7. The multivalued dependency E pp F holds on R if in any legal relation r(R).Multivalued Dependencies (MVDs) Let R be a relation schema and let E R and F R.
) Tabular representation of E pp F Database System Concepts 7.MVD (Cont. Korth and Sudarshan .56 ©Silberschatz.
z2. z1. Y. w2 > r then < y1. w2 > r and < y1. z2. w1 > r Note that since the behavior of Z and W are identical it follows that Y pp Z if Y pp W Database System Concepts 7. w1 > r and < y2. Korth and Sudarshan .57 ©Silberschatz. W We say that Y pp Z (Y multidetermines Z) if and only if for all possible relations r(R) < y1. z1. Z.Example Let R be a relation schema with a set of attributes that are partitioned into 3 nonempty subsets.
Database System Concepts 7.) In our example: course pp teacher course pp book The above formal definition is supposed to formalize the notion that given a particular value of Y (course) it has associated with it a set of values of Z (teacher) and a set of values of W (book). Korth and Sudarshan . Note: If Y p Z then Y pp Z Indeed we have (in above notation) Z1 = Z2 The claim follows.58 ©Silberschatz. and these two sets are in some sense independent of each other.Example (Cont.
59 ©Silberschatz. To specify constraints on the set of legal relations. To test relations to determine whether they are legal under a given set of functional and multivalued dependencies 2. We shall thus concern ourselves only with relations that satisfy a given set of functional and multivalued dependencies. Database System Concepts 7. If a relation r fails to satisfy a given multivalued dependency. Korth and Sudarshan . we can construct a relations rd that does satisfy the multivalued dependency by adding tuples to r.Use of Multivalued Dependencies We use multivalued dependencies in two ways: 1.
dependencies Database System Concepts 7. every functional dependency is also a multivalued dependency The closure D+ of D is the set of all functional and multivalued dependencies logically implied by D. it is better to reason about sets of using a system of inference rules (see Appendix C). which seem to be most common in practice For complex dependencies. We can compute D+ from D.60 ©Silberschatz. we can derive the following rule: If E p F. We can manage with such reasoning for very simple multivalued dependencies. then E pp F That is.Theory of MVDs From the definition of multivalued dependency. Korth and Sudarshan . using the formal definitions of functional dependencies and multivalued dependencies.
e.Fourth Normal Form A relation schema R is in 4NF with respect to a set D of functional and multivalued dependencies if for all multivalued dependencies in D+ of the form E pp F. where E R and F R. F E or E F = R) E is a superkey for schema R If a relation is in 4NF it is in BCNF Database System Concepts 7..61 ©Silberschatz. at least one of the following hold: E pp F is trivial (i. Korth and Sudarshan .
62 ©Silberschatz. Korth and Sudarshan .Restriction of Multivalued Dependencies The restriction of D to Ri is the set Di consisting of All functional dependencies in D+ that include only attributes of Ri All multivalued dependencies of the form E pp (F Ri) where E Ri and E pp F is in D+ Database System Concepts 7.
Note: each Ri is in 4NF. and decomposition is losslessjoin Database System Concepts 7. Let Di denote the restriction of D+ to Ri while (not done) if (there is a schema Ri in result that is not in 4NF) then begin let E pp F be a nontrivial multivalued dependency that holds on Ri such that E p Ri is not in Di.4NF Decomposition Algorithm result: = {R}. compute D+. and EF!J. end else done:= true.F) (E. done := false.63 ©Silberschatz. F). Korth and Sudarshan . result := (result .Ri) (Ri .
B) b) R2 = (A. G. G. C. I) c) R3 = (C. C. H. G. B. Korth and Sudarshan .64 ©Silberschatz. H. A pp I Database System Concepts 7. G) (R1 is in 4NF) (R2 is not in 4NF) (R3 is in 4NF) (R4 is not in 4NF) (R5 is in 4NF) (R6 is in 4NF) Since A pp B and B pp HI. I) f)R6 = (A. C.Example R =(A. A pp HI. H) d) R4 = (A. G. I) F ={ A pp B B pp HI CG pp H } R is not in 4NF since A pp B and A is not a superkey for R Decomposition a) R1 = (A. I) e) R5 = (A. C.
leads to a normal form called domainkey normal form. Problem with these generalized constraints: i hard to reason with. Korth and Sudarshan .Further Normal Forms join dependencies generalize multivalued dependencies lead to projectjoin normal form (PJNF) (also called fifth normal form) A class of even more general constraints. and no set of sound and complete set of inference rules.65 ©Silberschatz. Hence rarely used Database System Concepts 7.
which we then test/convert to normal form. Database System Concepts 7. Normalization breaks R into smaller relations.Overall Database Design Process We have assumed schema R is given R could have been generated when converting ER diagram to a set of tables. R could have been the result of some ad hoc design of relations. R could have been a single relation containing all attributes that are of interest (called universal relation). Korth and Sudarshan .66 ©Silberschatz.
employee entity with attributes departmentnumber and departmentaddress. the tables generated from the ER diagram should not need further normalization. Korth and Sudarshan . but rare .67 ©Silberschatz.ER Model and Normalization When an ER diagram is carefully designed. However.g. and an FD departmentnumber p departmentaddress Good design would have made department an entity FDs from nonkey attributes of a relationship set possible. in a real (imperfect) design there can be FDs from nonkey attributes of an entity to other attributes of the entity E. identifying all entities correctly.most relationships are binary Database System Concepts 7.
rn (Rn) be a set of relations A tuple r of the relation ri is a dangling tuple if r is not in the relation: Ri (r1 The relation r1 r2 « rn) r2 « rn is called a universal relation since it involves all the attributes in the ³universe´ defined by R1 R2 « Rn If dangling tuples are allowed in the database. r2 (R2).Universal Relation Approach Dangling tuples ± Tuples that ³disappear´ in computing a join. Database System Concepts 7. Korth and Sudarshan . Let r1 (R1)..68 ©Silberschatz. instead of decomposing a universal relation. we may prefer to synthesize a collection of normal form schemas from a given set of attributes. «.
customername) Universal relation would require null values. loannumber) (loannumber. and have dangling tuples Database System Concepts 7.g. They represent incomplete information E. Korth and Sudarshan .Universal Relation Approach Dangling tuples may occur in practical database applications. amount) (loannumber. may want to break up information about loans into: (branchname.69 ©Silberschatz.
g. branchname Reuse of attribute names is natural in SQL since relation names can be prefixed to disambiguate names Database System Concepts 7. customername.70 ©Silberschatz. Korth and Sudarshan . it can't be null either!) Universal relation requires unique attribute names unique role assumption e.) A particular decomposition defines a restricted form of incomplete information that is acceptable in our database. Above decomposition requires at least one of customername. amount in order to enter a loan number without using null values branchname or Rules out storing of customername.Universal Relation Approach (Contd. amount without an appropriate loannumber (since it is a key.
Denormalization for Performance
May want to use nonnormalized schema for performance E.g. displaying customername along with accountnumber and balance requires
join of account with depositor
Alternative 1: Use denormalized relation containing attributes of account as well
as depositor with all above attributes
faster lookup Extra space and extra execution time for updates extra coding work for programmer and possibility of error in extra code Alternative 2: use a materialized view defined as
account
depositor
Benefits and drawbacks same as above, except no extra coding work for programmer and avoids possible errors
Database System Concepts
7.71
©Silberschatz, Korth and Sudarshan
Other Design Issues
Some aspects of database design are not caught by normalization Examples of bad database design, to be avoided:
Instead of earnings(companyid, year, amount), use
earnings2000, earnings2001, earnings2002, etc., all on the schema (companyid, earnings). Above are in BCNF, but make querying across years difficult and needs new table each year companyyear(companyid, earnings2000, earnings2001, earnings2002) Also in BCNF, but also makes querying across years difficult and requires new attribute each year. Is an example of a crosstab, where values for one attribute become column names Used in spreadsheets, and in data analysis tools
Database System Concepts
7.72
©Silberschatz, Korth and Sudarshan
Proof of Correctness of 3NF Decomposition Algorithm
Korth and Sudarshan .74 ©Silberschatz.Correctness of 3NF Decomposition Algorithm 3NF decomposition algorithm is dependency preserving (since there is a relation for every FD in Fc) Decomposition is lossless join A candidate key (C) is in one of the relations Ri in decomposition Closure of candidate key under Fc must contain all attributes in R. Follow the steps of attribute closure algorithm to show there is only one tuple in the join result for each tuple in Ri Database System Concepts 7.
Database System Concepts 7. Consider each case separately.Correctness of 3NF Decomposition Algorithm (Contd.75 ©Silberschatz. (We need only consider FDs whose righthand side is a single attribute.) Claim: if a relation Ri is in the decomposition generated by the above algorithm.) Now. Korth and Sudarshan . then Ri satisfies 3NF. B can be in either F or E but not in both. Let Ri be generated from the dependency E pF Let K pF be any nontrivial functional dependency on Ri.
E must be contained in the attribute closure of K.Correctness of 3NF Decomposition (Contd. we can derive E pB (since K E F.) Case 1: If B in F: If K is a superkey. Database System Concepts 7. Now. Attribute closure not have used E pF . by using attribute closure on K. Thus.76 ©Silberschatz. since we assumed K is not a superkey. which is not possible.if it had been used. if B is in F then K must be a superkey. Korth and Sudarshan . and the second condition of 3NF must be satisfied. which is not possible since E pF is in Fc. the 2nd condition of 3NF is satisfied Otherwise E must contain some attribute not in K Since K p B is in F+ it must be derivable from Fc. B is extraneous in the righthand side of E pF. and F K since K p B is nontrivial) Then. using Ep (F.{B}) and K p B.
Korth and Sudarshan .77 ©Silberschatz. This shows exactly why the third alternative is present in the definition of 3NF. Database System Concepts 7.Correctness of 3NF Decomposition (Contd.) Case 2: B is in E. Q. we cannot show that K is a superkey. In fact. Since E is a candidate key.D. the third alternative in the definition of 3NF is trivially satisfied.E.
End of Chapter .
Korth and Sudarshan .79 ©Silberschatz.Sample lending Relation Database System Concepts 7.
Korth and Sudarshan .80 ©Silberschatz.Sample Relation r Database System Concepts 7.
Korth and Sudarshan .The customer Relation Database System Concepts 7.81 ©Silberschatz.
Korth and Sudarshan .The loan Relation Database System Concepts 7.82 ©Silberschatz.
Korth and Sudarshan .The branch Relation Database System Concepts 7.83 ©Silberschatz.
84 ©Silberschatz.The Relation branchcustomer branch Database System Concepts 7. Korth and Sudarshan .
Korth and Sudarshan .The Relation customerloan customer Database System Concepts 7.85 ©Silberschatz.
Korth and Sudarshan .86 ©Silberschatz.The Relation branchcustomer branch customerloan customer Database System Concepts 7.
Korth and Sudarshan .An Instance of Bankerschema Banker Database System Concepts 7.87 ©Silberschatz.
88 ©Silberschatz. Korth and Sudarshan .Tabular Representation of EppF Database System Concepts 7.
Korth and Sudarshan .89 ©Silberschatz.Relation bc: An Example of Reduncy in a BCNF Relation bc: Database System Concepts 7.
90 ©Silberschatz. Korth and Sudarshan .An Illegal bc Relation Database System Concepts 7.
Korth and Sudarshan .Decomposition of loaninfo loan Database System Concepts 7.91 ©Silberschatz.
Relation of Exercise 7.92 ©Silberschatz. Korth and Sudarshan .4 Database System Concepts 7.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.